SHAKE SK7 as an AHRS - Attitude and Azimuth Euler angle / Quaternion output demonstration

This is a demo of the SHAKE SK7 as an Attitude Heading Reference System (AHRS). It outputs Euler angle / Quaternions for it's attitude ( pitch and roll ) and azimuth (yaw) with respect to earth's North-East-Down reference plane. The algorithm operates on the SHAKEs internal DSP microcontroller and outputs the data over a Bluetooth serial port profile or USB. It uses Analog devices ADXRS610 MEMs Gyros, ST accelerometers and magnetometers. The visualization is written and runs in Matlab on a Windows XP PC.
More info on the SHAKE SK7 can be found here -
http://code.google.com/p/shake-drivers/

Category:

Science & Technology

Tags:

• IMU
• Inertial
• Sensing
• Gyro
• Angular
• rate
• accelerometer
• magnetometer
• DCM
• SHAKE
• SK7
• 'SAMH
• Engineering
• Services'
• SAMH
• Attitude
• Pitch
• Roll
• Yaw
• Azimuth
• Navigation
• 'Sensor
• Fusion'
• oreintation
• Quaternion
• AHRS

[hols2000]

Where can I buy one ?

[thepiper21]

@hols2000 Hi, you can order one by contacting me (my contact details are available along with all documentation of the SK7 on the SHAKE Drivers Google code webpage )

cheers, Stephen

[thepiper21]

Yes, from the DCM the Euler angles can be calculated. I use these equations -

ROLL = atan2( DCM32 , DCM33 )

PITCH = asin( -DCM31 )

YAW = atan2( DCM21 , DCM11 )

and the DCM elements can be calculate from the quaternion elements with these-

DCM11 = qA*qA + qB*qB + qC*qC - qD*qD

DCM21 = qB*qC + qA*qD

DCM31 = qB*qD - qA*qC

DCM32 = qC*qD + qA*qB

DCM33 = qA*qA - qB*qB + qC*qC + qD*qD

[lucasvago2]

Hello, can I ask you how do you do with quaternions to your cube rotate to any angle?? I trying to do a program like yours using quaternions, but when i try to rotate over the Y axis (pitch angle) I am stopped by a discontinuit given origined from the quaterions north and south pole. How can I revert that problem? Any suggestion? Thx

[thepiper21]

@lucasvago2 Hi,

The SK7 outputs quaternions to Matlab and in Matlab I perform the graphic rotations by first converting the quaternions into DCM rotation matrices and then it's just a simple matter of multiplying the graphic vectors by them.

[w8ryanb]

Awesome. Did you implement a Kalman filter? Would you mind sharing the code?

[thepiper21]

It uses a sensor fusion algorithm not a Kalman filter. Unfortunately we're not in a position to share the source code, sorry

[givechance87]

out of curiosity , how did you deal with the heading error caused by roll and pitch . Since I found that the error was huge even I tried to use those 2 common magnetometer compensation equations to get a more accurate measurement .Especially when the roll and pitch were approximate to 90 degree .

[aplejuce14]

second that on the code - trying to understand the filtering methods by reading code. its the only way i can.

thanks!

[lucasvago2]

Ok, but if you want the euler angles. Do you calculate them from the DCM matrix?

[w8ryanb]

No problem. Keep the good work.